Research On Low Altitude UAV's Passive Acoustic Detection And Location Method

At present,the UAV makes low altitude security more difficult.In the military,security monitoring and other fields,passive acoustic detection method has a wide range of applications.However,in the actual engineering application of low-altitude target detection,the traditional passive acoustic detection and location method has the problems of lower positioning accuracy under low signal-to-noise ratio.In response to the problems,four-rotor UAV is the research object in this paper.The focus is on the passive acoustic location algorithm based on the time-delay estimation and location algorithm of the three-dimensional seven-element acoustic sensor array.The improved maximum likehood-Phase transform-generalized cross correlation algorithm(ML-PHAT-GCC)is proposed to achieve the purpose of improving the positioning accuracy of time delay.To reduce the target distance,azimuth and pitch angle error,a dual-array model is established to research the improved artificial fish school algorithmthe unscented kalman partcle filter(ASFA-UPF)passive tracking algorithm.The main work of this paper is as follows:Firstly,the paper analyzes the time domain and frequency domain of the four-rotor UAV's acoustic signals,and analyzes the factors that affect the propagation of acoustic signals in the acoustic field.Due to the presence of a large number of environmental noise and reverberation noise in the sound field environment where the UAV is located,the high-frequency signals are removed by the band-pass filter,and the noise signal is reduced by LMS adaptive algorithm.Secondly,the positioning accuracy of planar five-element acoustic sensor array and three-dimensional seven-element acoustic sensor array are derived in this paper.Several kinds of weighted functions of generalized cross-correlation are analyzed.ML-PHAT-GCC algorithm is given.Up-sampling time delay estimation algorithm based on the rational multiple is given in this paper.The effects of different sampling frequencies on the distance,azimuth angle and elevation angle are simulated and analyzed.Thirdly,this paper presents a spatial cross-location algorithm for dual acoustic sensor arrays,giving the target state equation and observation equation.The artificial fish school algorithm is used to improve the unscented particle filter algorithm to overcome particle impoverishment problem.The performance of its tracking algorithm is compared with traditional particle filtering algorithm.Finally,the low altitude UAV passive acoustic detection experimental platform is designed and constructed,which includes acoustic sensor array module,multi-channel data acquisition module,data processing module and display module.Through the analysis of experimental data,the results show that the improved ML-PHAT-GCC delay estimation algorithm is significantly better than the traditional algorithm in this paper.Increasing the sampling frequency and the array element spacing can reduce the passive acoustic localization error.When the sampling frequency is 200 kHz,the distance error,azimuth angle error and pitch angle error of the improved algorithm are reduced by 2.477%,3.866% and 1.187% compared with the traditional algorithm.When the array element spacing is 0.65 m,the positioning error is the smallest,which is 9.198%,6.99% and 6.091% less than the positioning error when the array element spacing is 0.10 m.That verifies the effectiveness of the improved method.